Is synovial C-reactive protein a useful marker for periprosthetic joint infection?

BACKGROUND: Serum C-reactive protein (CRP) is a general marker of inflammation, and recent studies suggest that measurement of CRP in synovial fluid may be a more accurate method for diagnosing periprosthetic joint infection (PJI). QUESTIONS/PURPOSES: We aimed to (1) determine if there is a correlation between serum and synovial CRP values, (2) establish cutoff values for diagnosing infection based on serum and synovial CRP, and (3) compare the utility of measuring CRP in synovial fluid versus serum for the diagnosis of PJI using standard assay equipment available at most hospitals.
METHODS: Between February 2011 and March 2012, we invited all 150 patients scheduled for revision TKA (84) or THA (66) to participate in this prospective study, of whom 100% agreed. Data ultimately were missing for 31 patients, leaving 60 patients undergoing revision TKA and 59 undergoing revision THA (71% and 89% of the original group, respectively) for whom CRP level was measured in serum and synovial fluid samples. Patients were deemed to have a PJI (32) or no infection (87) using Musculoskeletal Infection Society criteria. Serum and synovial CRP levels were assayed using the same immunospectrophotometer and the correlation coefficient was calculated. Receiver operating characteristic curve analyses were performed to compare utility in diagnosing PJI, which included area under the curve, diagnostic threshold, and test sensitivity, specificity, predictive values, and accuracy. In 22 of 150 patients (14.7%), synovial CRP could not be measured because the sample was too viscous or hemolyzed.
RESULTS: In the analyzed 119 samples, there was a strong correlation (r = 0.76; p < 0.001) between synovial and serum CRP. The area under the curve was 0.90 both for the synovial fluid (95% CI, 0.82-0.97) and serum (95% CI, 0.84-0.96) CRP assays. The diagnostic thresholds were 6.6 mg/L for synovial fluid and 11.2 mg/L for serum. Sensitivities, specificities, positive predictive value, negative predictive value, and accuracies were similar for synovial fluid and serum assays.
CONCLUSIONS: Although recent studies have suggested a superiority of synovial fluid CRP over serum CRP for the diagnosis of PJI, we found that measurement of CRP in synovial fluid rather than serum using readily available assay equipment does not offer a diagnostic advantage in detection of PJIs. LEVEL OF EVIDENCE: Level II, diagnostic study. See the Instructions to Authors for a complete description of levels of evidence.